J . Am. Chem. SOC.1981, 103, 7389-7390
Table I. Quantum Yields of the Trans -+ Cis Photoisomerization of all-trans-Retinal as a Function of Temperature
Activation Barriers for the Trans Cis Photoisomerization of all- trans-Retinal +
@PIb
Walter H. Waddell* and Kohji Chihara
temp, solvent ethanol
Department of Chemistry, Carnegie-Mellon University Pittsburgh. Pennsylvania 15213 Received April 23, I981
-
7389
KO
trans .+ 9-cis 0.020
298 253 0.014 230 0.011 206 0.0052 183 0.0024 160 0.0018 3-methylpentane 298 0.015 0.10 0.079 273 253 0.0065 0.053 233 0.0086 0.041 213 0.0030 0.026 193 0.0014 0.018 173 0.00036 0.010 10-4-10-5 M in aerated solutions; @PI f 15%relative a + 2 K. to that measured at 298 K; he = 350 nm, 7-nm bandpass. 11cis-retinal is not a primary photoproduct in 3MP.
-
As the light-absorbing component of the visual protein rhodopsin and of bacteriorhodopsin, the cis trans and trans cis photoisomerizations of 11-cis- and all-trans-retinal (structure 1) have been widely studied since the retinyl chromophores of these proteins undergo such reactions following photon absorption.'
all-trans-r etinal (1 ) Experimental measurements of the quantum yields of photoisomerization of the isomeric retinals have been made upon direct excitation2" and triplet sen~itization,~.~ and primary photoproducts and product ratios have been determined.4s5-7-'0 Mechanistic details of these reactions have been based upon laser-induced kinetic and spectroscopic m e t h ~ d s , ~ , ~ Jand ' - ' ~upon theoretical calculations of the potential-energy surfaces of simple po1yenesl4-l7and retinylpolyenes.'8-20 Since only limited experimental information on the energetics of these excited-state isomerizations is available,2 but is necessary before a description of the photochemical and spectroscopic properties of the retinals can be complete, we have made quantitative investigations of the photoisomerization of all-trans-retinal as a function of temperature in a polar and a nonpolar solvent. Using high-pressure liquid chromatographic (LC) techniques the primary photoproducts and their relative ratios were determined at each temperature. The energy barrier of the excited-state isomerization of all-trans-retinal, including the barrier of each component isomerization, is calculated on the basis of a proposed isomerization mechanism. all-trans-Retinal (Sigma Chemical) was purified to >99% isomer pure by using high-pressure LC methods. 3-Methylpentane (3MP; Phillips Petroleum, 99.4%) was refluxed over and distilled from Dri-Na (Baker Chemical) immediately prior to use. Ethyl alcohol (EtOH; 200 proof, Publicker Industries) was distilled from calcium chloride immediately prior to use. Spectrophotometric
grade n-hexadecane (99+%), n-undecane (99%), and n-decane (99+%) were purchased from Aldrich Chemical and used as received. LC purifications and separations were accomplished by using the system and chromatographic conditions previously described.521 Room and low temperature absorption spectra were recorded on a Perkin-Elmer Model 575 spectrophotometer using long-stemmed, flat-faced 2-mm cells in conjunction with a flat-faced, doublejacketed liquid nitrogen Dewar. Temperatures between +25 and -180 OC (fl "C) were obtained by heating nitrogen vapors that were delivered from a reservoir of liquid nitrogen into the double-jacketed Dewar. Monochromatic radiation was obtained by using a Hanovia 1000-W Hg-Xe lamp and Bausch & Lomb High Intensity Grating Monochromator. was determined at room temperature by using the method we have previously e m p l ~ y e d . ~ ~ ~ ~ ~ ' ~ ~ Relative 4pI'swere determined between 0 and -125 OC by comparison to the room temperature dPIvalue obtained under otherwise identical experimental conditions. A correction for the difference in absorption intensity resulting from varying temperature was applied. All high-pressure LC analyses were performed at room temperature. All experimentation was performed under red lighting. Electronic absorption spectra of all-trans-retinal were recorded = between 25 and -125 OC; at room temperature in 3MP A, 368.5 nm, emax 50000 M-' cm-' and h (bandwidth at half-height) = 5300 cm-' and in EtOH A, = 382 nm, ,,e, 47 900 M-' cm-l and h = 5980 cm-'. With decreasing temperature for each solvent, A, undergoes a bathochromic shift and emax and h increase; however, the increase in emaxwith decreasing temperature has been shown to be due to a pure solvent contraction effect.23 Upon irradiation of all-trans-retinal in 3MP at room temperature, 13-cis- and 9-cis-retinal are formed as primary photoproducts in the relative ratio of 7:l; dP1= 0.12 f 0.02. 4pI obtained in n-decane, n-undecane, and n-hexadecane are somewhat lower (IO%), dPI= 0.10; however, these results are within experimental reproducibility of that measured in 3MP. In the polar solvent EtOH, 7-cis- and 11-cis-retinal are formed as primary photoproduct^;^.^ however, since 7-cis-retinal is present in only trace amounts, ca. 1% of the observed products, it was not analyzed. The 13-cis/l l-cis/9-cis product ratio was determined to be 11:4:1 which is similar to the ratios previously r e p ~ r t e d .&.~I = 0.33 in EtOH, a value notably higher than that measured in MeOH.4
(1) For a review, see: Rosenfeld, T.; Honig, B.; Ottolenghi, M.; Hurley, J.; Ebrey, T. G . Pure Appl. Chem. 1977, 49, 341. (2) Kropf, A.; Hubbard, R. Photochem. Photobiol. 1970, 12, 249. (3) Rosenfeld, T.;Alchalal, A.; Ottolenghi, M. J . Phys. Chem. 1974, 78, 336. (4) Waddell, W. H.; Crouch, R.; Nakanishi, K.; Turro, N. J. J. Am. Chem. SOC. 1976, 98, 4189. (5) Waddell, W. H.; Hopkins, D. L. J . Am. Chem. SOC.1977, 99, 6457. (6) Veyret, B.; Davis, S. G.; Yoshida, M.; Weiss, K. J. Am. Chem. SOC. 1978, 100, 3283. (7) Denny, M.; Liu, R. S.H. J. Am. Chem. SOC. 1977, 99, 4865. (8) Tsukida, K.; Kcdama, A.; Ito, M. J . Chromatogr. 1977, 134, 331. (9) Maeda, A.; Shichida, Y.; Yoshizawa, T. J . Biochem. 1978, 83, 661. (10) Sperling, W.; Carl, P.; Rafferty, C. N.; Dencher, N. A. Eiophys. Struct. Mech. 1977, 3, 79. ( 1 1 ) Bensasson, R.; Land, E. J. Nouu.J . Chim. 1978, 2, 503. (12) Menger, E. L.; Kliger, D. S. J. Am. Chem. SOC.1976, 98, 3975. (13) Hochstrasser, R. M.; Narva, D. L.; Nelson, A. C. Chem. Phys. Lett. 1976, 43, 15. (14) Warshel, A.; Karplus, M. J . Am. Chem. SOC.1972, 94, 5612. (15) Kushick, J. N.; Rice, S. A. J . Chem. Phys. 1976, 64, 1612. (16) Salem, L.; Bruckmann, P. Nature (London) 1975,258, 526. Salem, L. Acc. Chem. Res. 1979, 12, 87. (17) Ohmine, I.; Morokuma, K. J . Chem. Phys. 1980, 73, 1907. (18) Suzuki, H.; Nakashi, K.; Komatzu, T. J . Phys. SOC.Jpn. 1974, 37, 751. (19) Warshel, A.; Karplus, M. J . Am. Chem. SOC. 1974, 96, 5677. (20) Birge, R. R.; Schulten, K.; Karplus, M. Chem. Phys. Len. 1975,31, 451.
(21) Waddell, W. H.; West, J. L. J . Phys. Chem. 1980, 84, 134. (22) Waddell, W. H.; Dawson, P. M.; Hopkins, D. L.; Rach, K. L.; Uemura, M.; West, J. L. J. Liq.Chromatogr. 1979, 2, 1205. (23) Schaffer, A. M.; Waddell, W. H.; Becker, R. S. J . Am. Chem. SOC. 1974, 96, 2063.
0002-7863/81/1503-7389%01.25/0 0 1981 American Chemical Societv , , I
trans -, trans -, 11-cisc 13-cis 0.090 0.22 0.071 0.15 0.079 0.12 0.052 0.081 0.031 0.059 0.023 0.053
-
Communications to the Editor
7390 J. Am. Chem. Soc., Vol. 103, No. 24, 1981 Quantum yields relative to those measured at room temperature were determined to 160 K in EtOH and to 173 K in 3MP. Table I is a summary. Plots of log dP1vs. T I yield straight lines for each component isomerization in each solvent. In n-decane, the 4pIvalues and the relative product ratios measured at 298, 273, and 248 K were identical within experimental reproducibility. Activation barriers of the excited-state isomerization reactions can be obtained if the quantum yield data are treated as kinetic parametersz4 This can be accomplished by assuming a specific reaction pathway consistent with the data reported here and the following results obtained for all-trans-retinal: (a) triplet sensitization in a polar4 or nonpolar3 solvent yields no detectable isomerization (34pI< 0.003);4 (b) upon direct excitation only one carbon-arbon double bond is isomerized per p h o t ~ n ; (c) ~ * na~~ nosecond laser photolysis of all-trans-, 1 1-cis-, or 13-cis-retinal at room temperature gives rise to a distinct fluorescence spectrum (d) room temperature picosecond for each isomer (+F laser excitation reveals that the excited molecules instantaneously relax to two or more different singlet excited states ( T ~ 20 ps);13 (e) rotation of the retinoid skeleton by twists of 90' leads to a sudden polarization of the lowest singlet excited state;I6 (f) three excited singlet states are of comparable and on the basis of two-photon spectroscopy the Ag--like state is thought to be of lowest energy in EPA:6 while fluorescence studies in extensively dried 3MP conclude that the n,a* state is the lowest singlet excited
-
-
-
A pathway for the trans cis photoisomerization of alltramretinal that is based upon a rate-determining activated torsion to twisted states is presented in Scheme I. The essential features of Scheme I are excitation to the Bu-like excited singlet state, internal conversion to the photoreactive excited singlet state, torsion of the appropriate bond while the molecule is an excited singlet species (e.g., 9-cis-P*), and decay of this species to the ground-state cis or trans isomer. Using ki = Aie-EdRT,assuming that the twisted Scheme I trans
hv
kIC
trans* (Bu)
trans* (Bu-like)
trans* (Ag--like or n,a*)
-- ++ - + kD
trans*
trans* trans*
ks
kii
(1) (2)
trans
(3)
9-cis-P*
(4)
11-cis-P*
(5)
k13
trans* 13-cis-P* 9-cis-P* a 9-cis (1 - a) trans 11-cis-P* 13-cis-P*
a 11-cis
(1 - a) trans
a 13-cis
(1 - a) trans
(6)
(7) (8)
(9) states (e.g., 9-cis-P*) decay with equal probability to cis and trans double bonds (a = ai = 0.5), and using the approximation that (Ap-EifR3/(CAie-EifRT, Ai/CAi we can show that
-
In [(a/4i)
- (l/A)l
= In (k,/Ai)
+ Ei*/RT
(10)
(24) Saltiel, J.; Rousseau, A. D.; Sykes, A. J. Am. Chem. SOC.1972, 94, 5903. (25) For a recent review, see: Birge, R. R. Annu. Rev. Biophys. Bioeng. 1981, 10, 315. (26) Birge, R. R.; Bennet, J. A,; Fang, H. L-B.; Leroi, G. E. Springer Ser. Chem. Phys. 1978,3, 347-354. (27) Takemura, T.; Das, P. K.; Hug, G.; Becker, R. S. J . Am. Chem. Soc. 1976, 98,7099; 1978,100,2626. (28) Retinal fluorescence observed in 3MP (Waddell, W. H.; Schaffer, A. M.; Becker, R. S. J . Am. Chem. SOC.1973, 95, 8223.) is attributed to the presence of hydrogen-bonding species.*' That the photochemical properties of ull-trunr-retinal in 3MP is not being dominated by adventitious hydrogen-bonding species can be seen upon comparison of QpI and the number and ratio of primary photoproducts in 3MP and EtOH, which are notably different. In addition, it is thought that hydrogen-bonding does not measurably affect the intersystem crossing quantum yields measured in nonpolar solvents," since the equilibrium constant for hydrogen bonding is small at room temperat~re.~'
Table 11. Activation Barriers of the Excited-State Isomerization of all-trans-Retinal
producta isomer
solvent
E*, kcal/ @PIb
rz
mol
0.10 2.1 0.994 13-cis 0.015 3.2 0.944 9-cis 0.22 1.6 0.915 ethanol 134s 0.09 1.6 0.934 11-cis 0.02 2.5 0.966 94s a Determined by high-pressure liquid chromatographic methods. 10-4-105 M in aerated solutions;@pIf 15%;he = 350 nm, 7-nm bandpass. Coefficient of determination for the h e a r leastsquares fit of the data. 3-meth ylpentane
whereA = A i / C A i and can be calculated from the ratio of the room temperature quantum yields. Plots of the data according to eq 10 yield straight lines, the slopes of which yield Ei* values (Table 11). A correlation of the 4pIvalues of the component isomerization processes with the magnitude of its excited-state barrier (E*) may be possible. In 3MP the trans 13-cis isomerization process has a notably higher dPIvalue and measurably lower E* value than 9-cis process. Parallel results are obtained that of the trans in EtOH for these two component isomerization processes. One exception is that in EtOH the 4pIvalues of the trans 13-cis 11-cis processes are notably different; yet they and the trans have essentially identical E* values. Although not currently understood, this difference in behavior might be related to the fact that 11-cis-retinal exists as a temperature-dependent mixture of 12-s-cis and 12-s-trans conformer^;^^ hence the assumption that a l l = 0.5 may not be valid. Another complication is that a dependence of dPIon solvent viscosity is expected and while the values measured in n-decane, n-undecane, and n-hexadecane are approximately 10% lower than that measured in 3MP, these values are within experimental reproducibility. Although it is known that there is an increase in dispersive interactions between retinal and hydrocarbon solvents upon changing from 3 M P to n-decane to n - h e x a d e ~ a n e its , ~ ~effects upon the measured 4pI values is not u n d e r ~ t o o d . ~ ~ Finally, Scheme I may account for the notable differences in the photochemical properties ( c $and ~ ~the number and ratio of the primary photoproducts) of all-trans-retinal in 3MP and EtOH if photoisomerization were to occur from the Ag--like singlet excited state and intersystem crossing from the n,a* singlet excited state. In 3MP the n,r* is thought the lowest excited singlet state;" c $ = ~ 0.636J',31*32 and = 0.12. In ethanol, the Ag--like excited c 0.16*11932 and singlet state is thought lowest in energy;26 4 1 ~ = 4pI= 0.33. That photoisomerization is observed in 3 M P or intersystem crossing is observed in E t O H may be due to the proximity effects of the two states.33 We are indebted to Professor Jack Saltiel, Florida State University, for his helpful clarifications of the reaction kinetics and wish to thank the donors of the Petroleum Fund, administered by the American Chemical Society, for support of this research (Grant 12521-AC4).
-
-
-
-
(29) Becker, R. S.; Hug, G.; Das, P. K.; Schaffer, A. M.; Takemura, T.; Yamamoto, N.; Waddell, W. H. J. Phys. Chem. 1976,80, 2265. (30) An alternative mechanism can be proposed, one based upon the Bulike excited singlet state undergoing a temperature dependent, activated crossing to the photoreactive trans excited singlet state (reaction 2, Scheme I), with the subsequent torsional modes being temperature independent, e.&, E,, E,, Eg. Following this mechanism and the assumptions used to derive eq IO it can be shown that In [(f;a/Qi)- 11 = In ( k D / A )+ E * / R T (11)
- -
Plots of the data according to eq 11 do not reasonably fit one straight line for either solvent. Better fits of straight lines are obtained if component isomerizations are treated individually which is not consistent with the assumptions used to derive eq 1 1 . (31) Bensasson, R.; Land, E. J.; Truscott, T. J. Phorochem. Phorobiol. 1973, 17, 53. (32) Fisher, M. M.; Weiss, K. Photochem. Pholobiol. 1974, 20, 423. (33) For example, see: Madej, S. L.; Okajima, S.; Lim, E. C. J. Chem. Phys. 1976, 65, 1219. Wassam, W. A,; Lim, E. C. Ibid. 1978, 68, 433.
